
Matching methods for causal inference
Open Online Data Meetup - Vol. 8
Matching methods for causal inference
Open Online Data Meetup - Vol. 8
Wann: 2021-11-03, 19:00 - 20:00
Sprache: english
Registrierung: Hier
The Open Online Data Meetup (OODM) is back from summer break!
Join the 8th Open Online Data Meetup to learn more about matching methods for causal inference and get insights into a CorrelAid project.
The talks
Introduction to matching methods for causal inference without randomization
Randomizing people between groups that do or don’t receive an intervention or treatment allows us to measure the causal effect of that treatment. However, in many situations, randomization is not available because people choose for themselves to get treated, e.g. the effect of getting a product or service. In these cases, differences can arise between the treated and untreated people that are unrelated to the effect of the treatment itself. Matching methods often provide an effective and easy-to-interpret method for removing these confounding effects by balancing them across treated and untreated groups. Here the method is introduced with a focus on key considerations and practical implementation using the R package MatchIt (also callable in python).
Speaker: Garret O’Connell
CorrelAid Project Share out: The impact of Voting Advice Applications on local politics.
Voting advice applications (VAA) are web apps designed to match potentially hesitant voters with parties who share similar political positions. The users are usually required to complete an opinionnaire on current political affairs to commence the matching process. In Germany, VAA-operators mainly focused on party only matchings in their web applications prior to the European, Federal or State elections (i.e. Wahl-O-Mat). Local district elections were largely ignored because of the rather complicated candidate based electoral system. Therefore, a Stuttgart-based NPO “Team Tomorrow” developed “KOMUNAT”, an entirely candidate-based VAA for the local town elections in May 2019, to address the traditionally low voter turnout amongst the young first-time-voters. “Team Tomorrow” approached CorrelAid to gain insights into the statistics of the “KOMUNAT” app. Overall, two major topics had to be investigated: the user statistics for product improvement and the political analysis to assess both the political preferences of the app users and the potential impact of the VAA on local politics.
Speakers: Andreas Neumann and Aylin Kallmayer
Registration
Please register on Pretix.
How to access the Zoom Link: After registration on Pretix, you will receive an confirmation email containing a link to your order. When you open this link, you’ll find the Zoom link there.
The Open Online Data Meetup
The Open Online Data Meetup (OODM) is an online-based meetup series which provides the space to share interesting insights and entertaining stories from the field of data science with other people from the community. If you’re interested in speaking at the OODM, please write an email to the organizing team.